High Performance Computing (HPC)
High-performance computing encompasses advanced computation over parallel processing, enabling faster execution of highly compute intensive tasks such as climate research, molecular modeling, physical simulations, cryptanalysis, geophysical modeling, automotive and aerospace design, financial modeling, data mining and more. High-performance simulations require the most efficient compute platforms. The execution time of a given simulation depends upon many factors such as the number of CPU/GPU cores and their utilization factor and the interconnect performance, efficiency and scalability. Efficient high-performance computing systems require high-bandwidth, low-latency connections between hundreds or thousands of multi-processor nodes as well as high-speed storage systems.
One of HPC’s strengths is the ability to achieve best sustained performance by driving the CPU/GPU performance towards its limits. Compute clusters have become the most used hardware solution for HPC simulations due to their productivity and flexibility. Maximizing productivity in today’s HPC cluster platforms requires using enhanced data messaging techniques.
By providing low-latency, high-bandwidth, high message rate, transport offload for extremely low CPU overhead, Remote Direct Memory Access (RDMA) and advanced communications offloads, Mellanox interconnect solutions are the most deployed high-speed interconnect for large-scale simulations, replacing proprietary or low-performance solutions. Mellanox's Scalable HPC interconnect solutions are paving the road to Exascale computing by delivering the highest scalability, efficiency and performance for HPC systems today and in the future. Mellanox Scalable HPC solutions are proven and certified for a large variety of market segments, clustering topologies and environments (Linux, Windows). Mellanox is an active member of the HPC Advisory Council and is contributing to high-performance computing outreach and education around the world. Specific inquiries on Mellanox's Scalable HPC offering can be directed to HPC@mellanox.com.
High-performance computing enables a variety of markets and applications:
Bioscience data centers leverage HPC technology to reduce wait times and improve productivity. These data centers have deployed multi-core, multi-processor servers and high-speed storage systems, but they may be hamstrung by suboptimal performance of the network that links servers and storage.
EDA simulations often involve 3D modeling, fluid dynamics, and other compute-intensive processes that require high-performance computing (HPC) data center solutions.
Mechanical computer-aided design (MCAD) and computer-aided engineering (CAE) systems have adopted HPC cluster computing environments to speed processing times and reduce time to revenue for new products.
To reduce production lags, today’s media data centers invest in high-performance HPC cluster technology, combining the power of hundreds or thousands of CPUs in the service of a highly complex rendering task.
Oil and gas companies use HPC technology to minimize the time involved in processing massive amounts of data to reduce costs and speed time to production.
Meteorological forecasting and research requires high-speed processing of massive data inputs, so productivity depends on processing speed. Data centers used in weather research use HPC cluster technology, combining the power of hundreds or thousands of CPUs with large, high-speed storage systems.